During last few years density functional theory (DFT) based methods have been widely accepted by computational chemistry community as a reliably practical tool for the study of chemical reactions, especially, for large systems involving transition metal complexes. It is now a routine task to characterize reactants, products, and transition states for such large systems on the basis of DFT. However, relative little has been done to follow reaction path by the IRC scheme based on DFT or high level traditional ab initio methods. We have, for the first time, implemented the IRC method into DFT framework, and applied the combined DFT and IRC method to a wide range of chemical reactions involving main group (MG) and transition metal (TM) systems (see our publications). The MG chemical reactions include isomerization, substitution, abstraction, addition, and dissociation processes. The TM systems investigated are represented by the C-H and O-H bond activation of methanol by group-5 to group-8 do-metal oxo complexes. The IRC movie for the process of methanol oxidation to formaldehyde by chromium chloride is available.

The ongoing Project

We are currently studying the Brookhart catalytic systems for olefin polymerization process. These new catalysts are recently developed by Brookhart and co-workers and appear to be a promising alternative to the traditional Ziegler-Natta type as well as the new generation of the metallocene catalysts for ethylene and a-olefin polymerization. The Brookhart catalysts are palladium- and nickel-based complexes. The Pd(II) and Ni(II) initiator are cationic metal complexes that incorporate bulky diimine ligands. Brookhart et al. have shown that the palladium and nickel catalysts have very high activities, comparable those of metallocene catalysts used in some commercial polyolefin processes. Based on NMR studies, they proposed the catalytic polymerization mechanism as shown in Scheme 1.

 where the catalyst resting states are found to be alkyl olefin complexes, 1 and/or 3, and chain termination and generation of branched chain take place via a hydride olefin intermediate, 5.

We have being modeled the processes using both IRC and ab initio molecular dynamic simulations by Car-Parrinello-Projector-Augmented-Wave (CP-PAW) method within the nonlocal DFT formalism. Our preliminary results (for M=Ni, Ar=H, and R=CH3) show that the branched chain generated (2Æ5Æ6) via a kinetic process, in which the hydride complex 5 is the transition state like instead of an energy minimum point. The energy barrier (2Æ5) is about 12 kcal/mol. Both insertion and chain termination processes start from a strong pi-complex 3 which is about 20 kcal/mol lower in energy than 2 + ethylene and assigned to be the catalytic resting state. At the insertion transition state ethylene lies on the ring plane, the barrier is about 16 kcal/mol, whereas chain termination goes through a metastable hydride complex with the hydride lies on the plane and the entering ethylene and displaced polyethylene occupy two axial positions. The structures of the stationary points and barrier heights calculated by the conventional and CP-PAW DFT methods turn out to be in excellent agreement. Further dynamic details are being carried out. We are going to replace H by CH3 for Ar so as to investigation bulky ligand effects. We also want to compare the reaction paths generated by IRC and CP-PAW method, and use IRC as reaction coordinate for molecular dynamic simulations so to get more reliable dynamic pictures.

It has been the purpose of this project that (i) to microscopically understand the mechanism of ethylene polymerization by Brookhart catalysts, and (ii) to gain some insight into the critical factors that govern the rates of linear and branched chains as well as chain propagation and termination. The detailed results of this study are expected to submit for publication shortly.